基于基分类器自身的Bagging基分类器权重设置策略
首发时间:2019-03-18
摘要:Bagging算法是一种有效的提高分类任务中分类器性能的集成学习算法。文章提出一种Bagging算法中加权投票法权重的设置策略,该策略基于基分类器自身的性能去设置该基分类器在加权投票法中的权重。文章理论在图片分类数据集上进行验证,通过在图片分类数据集上的实验结果表明,文章中提出的加权投票法权重设置策略对于提高对分类器的分类准确率和泛化能力等各评价指标均有效,并且根据不同的基分类器性能指标来设置权重最终分类器的效果不同。
For information in English, please click here
Bagging base classifier weight setting strategy based on base classifier
Abstract:Bagging is an effective algorithm to improve the generalization ability of classifiers in classification tasks. This paper proposes a strategy for weight setting in the weighted voting method which for the bagging. The strategy sets the weight of the base learner in the weighted voting method based on the performance of the base classifier itself. The experimental results on the image classification dataset show that the strategy is effective for improving the classification accuracy and generalization ability of the classifier, and different weight are set according to the different performance of the base classifier, the final classifier has different effects.
Keywords: Software engineering ensemble learning Bagging weighted voting method
基金:
引用
No.****
同行评议
勘误表
基于基分类器自身的Bagging基分类器权重设置策略
评论
全部评论0/1000